Atrial fibrillation affects over 2 million people in the United States and accounts for 1% of all emergency department visits and nearly $7 billion in healthcare expenditures. Emergency department physicians have little information to guide the management of patients with symptomatic atrial fibrillation and, largely as a result, admit more than 65% for an inpatient stay. Studies suggest that nearly half of these admissions could be safely avoided. Emergency physicians need improved methods to identify the low risk atrial fibrillation patient who does not require hospitalization. We hypothesize that readily available emergency department data can be utilized in an atrial fibrillation clinical prediction rule to identify those patients at low or high risk for adverse outcome. We propose developing and validating a multivariable clinical prediction rule that accurately estimates risk for adverse outcome in emergency department patients with atrial fibrillation. We will describe and analyze the rule's utility by the distribution of assigned risk and its impact on physician decision-making. To develop our prediction rule, we will measure nineteen independent variables that are routinely available to the emergency physician within the first two hours of emergency department management. These variables will include patient history and physical examination findings, vital signs, and common laboratory studies. These data will be incorporated into the development of our prediction rule, while adhering to established clinical and biostatistical standards. Our proposed rule will predict risk for both 5-day and 30-day adverse outcomes. This 5-day outcome is more applicable to the emergency department setting where standard 30-day outcomes are highly dependent on unpredictable, post-visit patient and provider behavior. This proposed study will also collect and store blood and plasma samples for future analysis. The development of a highly accurate prediction rule will significantly advance the treatment of atrial fibrillation and reduce unnecessary hospitalizations. This is the initial step toward our goal of improving the emergency department management of atrial fibrillation and reducing unnecessary resource utilization through more appropriate dispositions and individualized pharmacologic treatment. Atrial fibrillation, a disorder in which the heart beats irregularly, affects over two million Americans and is associated with a four-fold increase in the risk of stroke. Nearly 75% of patients treated in emergency departments for atrial fibrillation are hospitalized and the annual treatment costs exceed $6.65 billion. Our clinical study will enhance management of patients with acute atrial fibrillation by improving physician disposition decisions and reducing unnecessary costly hospitalizations.